Jiamiao Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Jiamiao Yang has a total of 120 co-authors. Here are the top co-authors by number of shared publications:
- Yuan Qu
Shanghai Jiao Tong Univ, Sch Elect Informat & Elect Engn, Shanghai 200240, Peoples R China
21 shared publications Rongjun Shao
School of Automation and Intelligent Sensing, Shanghai Jiao Tong Universityhttps://ror.org/0220qvk04, Shanghai 200240, China
16 shared publications- Chunxu Ding
Shanghai Jiao Tong Univ, Sch Elect Informat & Elect Engn, Shanghai 200240, Peoples R China
15 shared publications - Qiaozhi He
Shanghai Jiao Tong University, Institute of Marine Equipment, Shanghai, China
13 shared publications - Rongjun Shao
School of Automation and Intelligent Sensing, Shanghai Jiao Tong University, Shanghai 200240, People’s Republic of China
11 shared publications Linxian Liu
10 shared publications
Latest Publications
- 2 November 2026
- Published by Optica Publishing Group in Optics Letters
- Vol. 51 (22) , 6159-6162
Continuous 3D microscopic measurement via spatial-resolution compensation
Extended-depth-of-field (EDOF) microscopy estimates surface height from axial focal positions, but elongation of the three-dimensional point-spread function makes axial resolution poorer than lateral resolution. We propose binocular oblique-view spatial-compensation microscopy (BOSCM), which uses two fixed, complementary oblique views to convert continuous lateral sample motion into simultaneous axial focal-plane scanning and lateral translation of the field of view. This geometry synchronizes focus-stack acquisition with large-area scanning and projects axial height variations onto the higher-resolution lateral direction. Global spatial calibration, constrained low-frequency residual compensation, and reliability-weighted fusion jointly recover dual-view height data and all-in-focus texture. Experiments show that BOSCM reduces spatial-resolution anisotropy, decreases the mean height RMSE by approximately 50%, improves the height-discrimination capability by approximately 18-fold, and improves reconstruction completeness on complex surfaces, enabling continuous large-area three-dimensional microscopic measurement.
- 2 November 2026
- Published by Optica Publishing Group in Optics Letters
- Vol. 51 (22) , 6159-6162
Continuous 3D microscopic measurement via spatial-resolution compensation
Extended-depth-of-field (EDOF) microscopy estimates surface height from axial focal positions, but elongation of the three-dimensional point-spread function makes axial resolution poorer than lateral resolution. We propose binocular oblique-view spatial-compensation microscopy (BOSCM), which uses two fixed, complementary oblique views to convert continuous lateral sample motion into simultaneous axial focal-plane scanning and lateral translation of the field of view. This geometry synchronizes focus-stack acquisition with large-area scanning and projects axial height variations onto the higher-resolution lateral direction. Global spatial calibration, constrained low-frequency residual compensation, and reliability-weighted fusion jointly recover dual-view height data and all-in-focus texture. Experiments show that BOSCM reduces spatial-resolution anisotropy, decreases the mean height RMSE by approximately 50%, improves the height-discrimination capability by approximately 18-fold, and improves reconstruction completeness on complex surfaces, enabling continuous large-area three-dimensional microscopic measurement.
- 23 June 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 203
Adaptive fringe density deflectometry measuring surfaces with large height fluctuation
- 23 June 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 203
Adaptive fringe density deflectometry measuring surfaces with large height fluctuation
- 15 May 2026
- Published by Optica Publishing Group in Optics Express
- Vol. 34 (10) , 19364-19381
Robust 3D topography measurement of large-range high-aspect-ratio structures based on dual-domain statistical filtering in SD-OCT
Non-destructive topography measurement of high aspect ratio (HAR) microstructures (e.g., deep micro-holes) over a large dynamic range remains a major challenge in precision manufacturing. While spectral-domain optical coherence tomography (SD-OCT) offers a promising non-contact solution, its application is severely hindered by the complex background and noise inherent to HAR structures. To address this, this study proposes a dual-domain statistical filtering (DDSF) reconstruction method combined with a conventional longitudinal scanning (Z-scan) architecture. Utilizing the statistical features of interference signals, this method achieves the robust extraction of physical interfaces under extreme conditions through wavenumber-domain principal component analysis filtering and spatial-domain Bayesian filtering. DDSF successfully reconstructs the 3D topography of a highly reflective metal micro-hole with an aspect ratio exceeding 8:1, without complex hardware system modifications. Overall, DDSF effectively extends the geometric detection limits of traditional reflective SD-OCT and holds the potential to become a highly reliable, large dynamic range topography measurement tool for industrial HAR structures.
- 15 May 2026
- Published by Optica Publishing Group in Optics Express
- Vol. 34 (10) , 19364-19381
Robust 3D topography measurement of large-range high-aspect-ratio structures based on dual-domain statistical filtering in SD-OCT
Non-destructive topography measurement of high aspect ratio (HAR) microstructures (e.g., deep micro-holes) over a large dynamic range remains a major challenge in precision manufacturing. While spectral-domain optical coherence tomography (SD-OCT) offers a promising non-contact solution, its application is severely hindered by the complex background and noise inherent to HAR structures. To address this, this study proposes a dual-domain statistical filtering (DDSF) reconstruction method combined with a conventional longitudinal scanning (Z-scan) architecture. Utilizing the statistical features of interference signals, this method achieves the robust extraction of physical interfaces under extreme conditions through wavenumber-domain principal component analysis filtering and spatial-domain Bayesian filtering. DDSF successfully reconstructs the 3D topography of a highly reflective metal micro-hole with an aspect ratio exceeding 8:1, without complex hardware system modifications. Overall, DDSF effectively extends the geometric detection limits of traditional reflective SD-OCT and holds the potential to become a highly reliable, large dynamic range topography measurement tool for industrial HAR structures.
- 4 May 2026
- Published by AIP Publishing in Applied Physics Letters
- Vol. 128 (18)
Unlimited field-of-view focus stacking microscopy based on tilted imaging
Focus stacking microscopy achieves high-resolution, all-in-focus imaging through multi-focus image fusion and is widely applied in fields such as biology, materials science, and semiconductor inspection. However, for large-scale samples, conventional methods require pausing at each depth-of-field imaging position to capture a multi-focus image stack, which introduces mechanical backlash errors, reducing inspection efficiency and stitching accuracy, and thus fails to meet the requirements for all-in-focus, large field-of-view (FOV), and high-throughput inspection. Here, we propose an unlimited FOV focus stacking microscopy method based on tilted imaging (TI-FSM). By tilting the microscopy system, we decompose the sample's one-dimensional displacement into motion along and perpendicular to the optical axis, where the former is used to acquire a multi-focus image stack for all-in-focus scanning, while the latter enables unlimited FOV expansion. Additionally, we employ a joint texture–topography reconstruction algorithm to achieve simultaneous reconstruction of texture and 3D topography. Experimental results demonstrate that using a microscopy system with an NA of 0.28, TI-FSM achieves a vertical reconstruction accuracy of 16.2 μm. It enables unlimited FOV expansion and exhibits strong all-in-focus imaging and 3D topography measurement capabilities for large-scale, complex-structured surfaces.
- 4 May 2026
- Published by AIP Publishing in Applied Physics Letters
- Vol. 128 (18)
Unlimited field-of-view focus stacking microscopy based on tilted imaging
Focus stacking microscopy achieves high-resolution, all-in-focus imaging through multi-focus image fusion and is widely applied in fields such as biology, materials science, and semiconductor inspection. However, for large-scale samples, conventional methods require pausing at each depth-of-field imaging position to capture a multi-focus image stack, which introduces mechanical backlash errors, reducing inspection efficiency and stitching accuracy, and thus fails to meet the requirements for all-in-focus, large field-of-view (FOV), and high-throughput inspection. Here, we propose an unlimited FOV focus stacking microscopy method based on tilted imaging (TI-FSM). By tilting the microscopy system, we decompose the sample's one-dimensional displacement into motion along and perpendicular to the optical axis, where the former is used to acquire a multi-focus image stack for all-in-focus scanning, while the latter enables unlimited FOV expansion. Additionally, we employ a joint texture–topography reconstruction algorithm to achieve simultaneous reconstruction of texture and 3D topography. Experimental results demonstrate that using a microscopy system with an NA of 0.28, TI-FSM achieves a vertical reconstruction accuracy of 16.2 μm. It enables unlimited FOV expansion and exhibits strong all-in-focus imaging and 3D topography measurement capabilities for large-scale, complex-structured surfaces.
- Open Access
- 15 April 2026
- Published by Springer Nature in Light: Science & Applications
- Vol. 15 (1) , 199
Large dynamic range Shack-Hartmann wavefront sensing based on a graph-theoretic computational model
The Shack-Hartmann wavefront sensor (SHWS) is a widely used non-interferometric wavefront measurement technique. However, for high-slope wavefronts, spot crosstalk and asymmetric distortion cause severe matching ambiguity and centroiding errors. This creates an inherent conflict between dynamic range and reconstruction accuracy. To address this, a graph-theoretic computational model named G-SHWS is proposed. By minimizing the global pairing cost of a bipartite graph constructed between fitted and actual spots, G-SHWS drives the fitted distribution to approximate the true distribution and maps the subaperture attribution of the fitted spots to the actual spots, achieving precise spot-subaperture matching under severe aliasing. Furthermore, incorporating a Graph Attention Network (GAT) embedded with SHWS matching topology, the model utilizes a graph structure to explicitly encode the matching relationships obtained from the matching process, and combines the spatial features and intensity morphology of spots to achieve high-precision reconstruction of strongly distorted wavefronts, effectively circumventing the inherent centroiding errors under large aberrations. Experimental results demonstrate that G-SHWS extends the measurable range of SHWS to 21 times the conventional limit while maintaining a reconstruction error of less than $$0.05{\rm{\lambda }}$$ , and remains robust under severe spot loss. These advancements significantly enhance the SHWS’s ability to measure complex aberrations, providing a reliable computational framework for large dynamic range wavefront sensing.- 15 April 2026
- Published by Springer Nature in Light: Science & Applications
- Vol. 15 (1) , 199
Open AccessLarge dynamic range Shack-Hartmann wavefront sensing based on a graph-theoretic computational model
The Shack-Hartmann wavefront sensor (SHWS) is a widely used non-interferometric wavefront measurement technique. However, for high-slope wavefronts, spot crosstalk and asymmetric distortion cause severe matching ambiguity and centroiding errors. This creates an inherent conflict between dynamic range and reconstruction accuracy. To address this, a graph-theoretic computational model named G-SHWS is proposed. By minimizing the global pairing cost of a bipartite graph constructed between fitted and actual spots, G-SHWS drives the fitted distribution to approximate the true distribution and maps the subaperture attribution of the fitted spots to the actual spots, achieving precise spot-subaperture matching under severe aliasing. Furthermore, incorporating a Graph Attention Network (GAT) embedded with SHWS matching topology, the model utilizes a graph structure to explicitly encode the matching relationships obtained from the matching process, and combines the spatial features and intensity morphology of spots to achieve high-precision reconstruction of strongly distorted wavefronts, effectively circumventing the inherent centroiding errors under large aberrations. Experimental results demonstrate that G-SHWS extends the measurable range of SHWS to 21 times the conventional limit while maintaining a reconstruction error of less than $$0.05{\rm{\lambda }}$$ , and remains robust under severe spot loss. These advancements significantly enhance the SHWS’s ability to measure complex aberrations, providing a reliable computational framework for large dynamic range wavefront sensing. - Open Access
- 1 January 2026
- Published by Journal of Applied Optics in Journal of Applied Optics
- Vol. 47 (2) , 349-356
基于混合视觉的金刚线布局智能检测方法
硅片切割工艺在太阳能电池制造等工业领域中具有关键地位,其中,金刚线线槽及线网布局的精确检测对硅片质量与切割精度至关重要。然而,传统人工检测方法存在效率低、一致性差的缺点,而纯深度学习方法在应对此类高密度、弱对比度的周期性结构时,常因特征混淆导致定位与分割精度不足。因此,提出了一种混合检测方法,该方法首先通过频域分析与曲线拟合构建线槽位置编码模型,实现线槽的微米级分割与定位,随后设计了一种轻量化网络DARMobileV2实现单一线槽图像的快速分类,最后结合位置编码信息重构出完整的线网布局情况。实验结果表明,在典型检测条件下,该方法在边缘计算设备上检测准确率不低于99.6%,检测时间低于1 s,相较传统的人工检测方法,错检率降低97.74%,速度提升52.4倍,为硅片精密切割工艺提供了一种高效、可靠的自动化检测解决方案。
- 1 January 2026
- Published by Journal of Applied Optics in Journal of Applied Optics
- Vol. 47 (2) , 349-356
Open Access基于混合视觉的金刚线布局智能检测方法
硅片切割工艺在太阳能电池制造等工业领域中具有关键地位,其中,金刚线线槽及线网布局的精确检测对硅片质量与切割精度至关重要。然而,传统人工检测方法存在效率低、一致性差的缺点,而纯深度学习方法在应对此类高密度、弱对比度的周期性结构时,常因特征混淆导致定位与分割精度不足。因此,提出了一种混合检测方法,该方法首先通过频域分析与曲线拟合构建线槽位置编码模型,实现线槽的微米级分割与定位,随后设计了一种轻量化网络DARMobileV2实现单一线槽图像的快速分类,最后结合位置编码信息重构出完整的线网布局情况。实验结果表明,在典型检测条件下,该方法在边缘计算设备上检测准确率不低于99.6%,检测时间低于1 s,相较传统的人工检测方法,错检率降低97.74%,速度提升52.4倍,为硅片精密切割工艺提供了一种高效、可靠的自动化检测解决方案。
- 7 November 2025
- Published by Optica Publishing Group in Optics Express
- Vol. 33 (23) , 48393-48402
Genetic-algorithm-based nonlinear calibration method for the digital micromirror device
As an important modulation device for optical imaging, precision detection, and high-end display, the nonlinear error of the digital micromirror device (DMD) has become a key performance-limiting factor. Traditional calibration methods rely on simplified models, struggle to characterize full-range nonlinearity, and cannot directly optimize the key parameter, like micromirror flipping time. To address these challenges, this paper proposes a genetic algorithm (GA)-based nonlinear calibration method. By using a photodetector (PD) for real-time light intensity feedback and GA to optimize the flipping time, high-precision calibration is achieved. Simulations and experiments demonstrate global optimization and uniform modulation, with DMD linearity improved by 42.2% post-calibration. Optimal parameters show strong adaptability across frame rates, enhancing optimization efficiency. Notably, the method remains effective under hardware-limited high frame rates, expanding the working range and applicability.
- 7 November 2025
- Published by Optica Publishing Group in Optics Express
- Vol. 33 (23) , 48393-48402
Genetic-algorithm-based nonlinear calibration method for the digital micromirror device
As an important modulation device for optical imaging, precision detection, and high-end display, the nonlinear error of the digital micromirror device (DMD) has become a key performance-limiting factor. Traditional calibration methods rely on simplified models, struggle to characterize full-range nonlinearity, and cannot directly optimize the key parameter, like micromirror flipping time. To address these challenges, this paper proposes a genetic algorithm (GA)-based nonlinear calibration method. By using a photodetector (PD) for real-time light intensity feedback and GA to optimize the flipping time, high-precision calibration is achieved. Simulations and experiments demonstrate global optimization and uniform modulation, with DMD linearity improved by 42.2% post-calibration. Optimal parameters show strong adaptability across frame rates, enhancing optimization efficiency. Notably, the method remains effective under hardware-limited high frame rates, expanding the working range and applicability.
- 26 September 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (9) , 095415
A fast optical design method for the objective under Köhler illumination to reduce ghost images
In high-precision microscopy, the Köhler illumination is often used to ensure clear surface details. However, it can introduce artifacts, such as ghost images, which must be effectively controlled. Here, we propose a matrix optics method (MOM) to estimate the ghost images formed in the microscope objective under the Köhler illumination. On the basis of MOM, we develop a function for the evaluation of image quality and ghost image intensity. Using this function, we can optimize the objective to mitigate the ghost images and preserve a good image quality simultaneously. The research results show that our method can reduce the intensity of ghost image by 60.6% as well as the time consumption from 158 s to 3 s for each iteration. This method can be potentially useful for the optical design of microscope objectives.
- 26 September 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (9) , 095415
A fast optical design method for the objective under Köhler illumination to reduce ghost images
In high-precision microscopy, the Köhler illumination is often used to ensure clear surface details. However, it can introduce artifacts, such as ghost images, which must be effectively controlled. Here, we propose a matrix optics method (MOM) to estimate the ghost images formed in the microscope objective under the Köhler illumination. On the basis of MOM, we develop a function for the evaluation of image quality and ghost image intensity. Using this function, we can optimize the objective to mitigate the ghost images and preserve a good image quality simultaneously. The research results show that our method can reduce the intensity of ghost image by 60.6% as well as the time consumption from 158 s to 3 s for each iteration. This method can be potentially useful for the optical design of microscope objectives.
- 15 September 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 192
UMSPU: Universal multi-size phase unwrapping network via mutual self-distillation and adaptive boosting ensemble segmenter
- 15 September 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 192
UMSPU: Universal multi-size phase unwrapping network via mutual self-distillation and adaptive boosting ensemble segmenter
- 28 May 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 190
Wavelet-based propagation model for the light modulated by a digital micromirror device
- 28 May 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 190
Wavelet-based propagation model for the light modulated by a digital micromirror device